Risk factors for the adverse cardiac outcomes in the patients with Takayasu’s arteritis.

Background : This study was aimed to investigate the long-term cardiac outcomes and find out risk factors for the deterioration of cardiac in with (TAK). Methods : One hundred and ninety-eight TAK patients were recruited in the Department of Rheumatology, Zhongshan Hospital, Fudan University from January 2012 to December 2017. The endpoint of observation was December 31, 2019. Patients were divided into normal, mild-moderate and severe subgroups based on the severity of structural and functional impairment of heart. Comparison of baseline characteristics and treatment among these three subgroups was conducted to identify the risk factors for adverse cardiac outcomes. Results : A total of 60.1% (119/198) of TAK patients exhibited heart involvement on echocardiography at baseline. 70.6% (84/119) and 79.8% (95/119) cases showed pathological valvular and atrioventricular abnormalities, respectively. 68.7% (136/198) of patients were followed up and the median follow-up duration was 28.6 (12.3-43.4) months. 42.7% (58/136) and 22.8% (31/136) of patients had progress in heart’s structure and function. Cox regression showed that serum TNF-α level (HR=1.028, p= 0.021), aortic regurgitation (HR=3.109, p<0.001) and ventricular hypertrophy (HR=2.090, p= 0.006) at baseline were risk factors for the deterioration of cardiac structure. The cases with serum TNF-α levels (≥8.1pg/ml), mild aortic regurgitation and ventricular hypertrophy at baseline had highest risk of structural progression at 83.4%. Using normal subgroup as a control, the deterioration of cardiac structure and function was the most severe in the severe subgroup, with hazard ratios of 16.2 (95% CI, 6.0-44.1, p < 0.001) and 9.0 (95% CI, 3.3-24.5, p<0.001), respectively. The deteriorating events typically occurred within 2 years after diagnosis. Conclusions


Introduction
Takayasu's arteritis (TAK) is a typically systematic large vessel vasculitis that is characterized by progressive inflammation and fibrosis in the wall of the aorta and its major branches [1,2]. TAK has apparently ethnic associations, and young Asian women are the predominant patient population [3].
Epidemiologic researches show that TAK has a low frequency of death, but a high risk of organ dysfunction; hence, TAK patients often experience low quality of life over a considerably long period [4]. Heart is one of the important organs commonly attacked in TAK, but the subsequent changes of cardiac involvement remain unclear.
The involvement of heart has been reported in 8.6%-39.9% of TAK patients [5,6]. TAK-related heart diseases mainly include valvular insufficiency or stenosis, hypertensive heart disease, ischemic cardiomyopathy and so on [7][8][9][10][11]. Different degrees of heart failure occurred frequently in these TAK subgroup patients, which results in poor prognosis even death [12]. However, the majority of recent studies on the TAK-related heart diseases are case reports, case series, or cross-sectional studies, and it lacks large sample, prospective and dynamic observational studies to determine the development of cardiac involvement and its risk factors. Echocardiography is a widely used noninvasive tool to evaluate heart structural disease and cardiac dysfunction in TAK patients [13,14].
We therefore designed this prospective study to investigate the characteristics of TAK patients with cardiac involvement and clarify the underlying risk factors for the adverse cardiac outcomes occurring during long-term follow-up. Kerr score proposed by the National Institutes of Health (NIH) was used to assess the disease activity.

Patients
Patients who complained of serious infection, tumor, and current or past history of viral myocarditis or other rheumatic diseases were excluded. This study has been approved by the Ethics Committee of Zhongshan Hospital, Fudan University [B2013-115(3)], and written informed consent was obtained from all the participants.

Clinical database
The clinical records including demographic data (gender, age, etc.), clinical manifestation, laboratory test results, and imaging findings were electronically stored in a database managed by a professional team. Whole-body magnetic resonance angiography (MRA) was used instead of digital subtraction angiography (DSA) for the diagnosis and vascular evaluation of TAK. The MRA parameters were referred to our previous reports [15]. Echocardiography was administrated in the Philips iE33 system (Philips Medical Systems, Bothell, WA, USA), and assessed by two independent echocardiographic experts.

Follow-up
The endpoint of observation was December 31, 2019. For naive or active patients, the visit was recommended once a month during the induction period and once per 3-month during the stable period. At each visit, the symptoms and signs were collected, and peripheral blood was extracted to determine the routine full blood count, erythrocyte sedimentation rate (ESR), hypersensitive Creactive protein (hsCRP) and inflammatory factors such as interleukin-6 (IL-6) and tumor necrosis factor-α (TNF-α). Echocardiography was performed during follow-up based on the condition of the TAK patients, or at an interval of 6 months.

Treatment
TAK treatment includes the induction and stable period. Glucocorticoid is the first-line drug for TAK.
The initial dose of prednisone or its equivalents was oral 40-60 mg per day and tapered to 10-15mg in ) is essential to achieve remission and prevent relapse. The major treatment in the induction period included intravenous CTX 0.4-0.5g/m 2 per month, oral LEF 10-20mg per day and myocophenolate mofetil (MMF) 1.0-1.5g per day. After patients achieved clinical remission, intravenous CTX was always replaced by oral CTX or LEF in the stable period, while those oral induction drugs kept up but reduce their doses according to the condition of disease. To improve the symptoms of ischemia and maintain organ function, some patients will undergo surgical intervention. The cardiac surgery procedures include aortic valve replacement, mitral valvuloplasty, pulmonary artery stent implantation, coronary artery bypass surgery, and percutaneous transluminal angioplasty.

Subgrouping based on the severity of cardiac involvement
Patients in severe group were those with severe valvular or atrioventricular lesions, heart function classified as NYHA III or IV, or both, and severe pulmonary hypertension with right heart structural changes and NYHA III or IV. Patients in mild-moderate group were those with mild to moderate valvular lesions, cardiac structural changes, and pulmonary hypertension with NYHA I or II. The other patients were considered as normal.

Definition of outcomes of cardiac structure and function
The deterioration of cardiac structure was considered as new lesions or significant progression of structural parameters occurred, and the deterioration of cardiac function was defined as one grade of NYHA increased as described previously [16,17]. The improvement of cardiac structure was defined as a decreased cardiac chamber, thinned septum or reduced valvular regurgitation on the echocardiography. Patients with clinical remission and reduced NYHA grade were defined as improvement in cardiac function. The others were considered stable or in maintenance.

Statistical analysis
Continuous variables were described as median (Q1, Q3), while categorical variables were summarized as percentages. The Kruskal-Wallis test or Chi-square test was performed to compare the differences of baseline features among different subgroups. Variables that with P<0.15 in univariate analysis were chosen to perform Cox proportional-hazards regression multivariate models. Further, the risk factors that entered into the Cox regression models were also included in the matrix-based logistic regression to assess the probability of deterioration of cardiac structure and function. Kaplan Meier curve with log-rank test was applied to compare differences of free probability in the adverse outcomes of cardiac structure and function between subgroups. Statistical significance was set at p<0.05 (two-sides). Analysis was performed using SPSS software (version25.0; IBM Crop, Armonk, NY,

Baseline clinical characteristics of TAK patients
A total of 198 TAK patients who underwent echocardiography were enrolled in our study ( Table 1).
The male/female ratio was 1:3.6 (43 men, 155 women), and the median age of onset was 28.9(21.8-42.0) years. The median disease duration at first echocardiography was 16

Baseline echocardiographic characteristics of TAK patients
The general echocardiographic index is shown in Table 2

Comparison of the severity of cardiac structure and function impairment among subgroups
The clinical characteristics of 3 subgroups are shown in Table 1. At baseline, the onset age (p<0.001), NYHA grade (p<0.001) and the rate of hypertension (p=0.035) were significantly different among 3 subgroups. The frequency of cardiac dysfunction symptoms in the severe subgroup was markedly greater than those in the other two subgroups (all p<0.05). There were no significant differences in gender, TAK classification or NIH score (p>0.05). With respect to the laboratory test, the acute reactants like ESR, hsCRP, and proinflammatory cytokines such as TNF-α and IL-6 levels were elevated in TAK patients, although the differences across three subgroups were not significant (all p>0.05).

Takayasu's arteritis-related treatment
The medical treatments administered to TAK patients are described in Table 3. During the induced remission period, the median initial prednisone dose was 40 mg/day without difference between subgroups. There was no significant difference in the doses of anti-rheumatic drugs among 3 subgroups (p>0.05) except leflunomide (p=0.041). In terms of heart disease treatment, the largest number of angiotensin converting enzyme inhibitors (ACEI), β-receptor blockers, and digoxin, diuretic, and antilipemic drugs were used in the severe subgroup (p<0.05).
Ten cases (7 cases in severe and 3 cases in mild-moderate group) underwent cardiac surgery and intervention, including aortic valve replacement, Bentall procedure, mitral valvuloplasty, pulmonary or coronary stent implantation and coronary artery bypass grafting. Additionally, divided into two groups according to whether cardiac structure deteriorated, there was no significant difference in the doses of anti-rheumatic drugs among two groups. But more doses of ARB, β-receptor blockers, and diuretic drugs were used in the worse group (p<0.05). There were 9 patients treated with TNF-α antibody, 20% (1/5) of the 5 patients who were followed up had cardiac structure and function deterioration, and another 20% had improved cardiac structure and function.
Of the 12 patients treated with Tocilizumab, 2 cases (16.7%) had cardiac structure and function deterioration, one case 8.3% had improved cardiac structure and function.
Using the normal subgroup as a control, the survival curve ( Figure 5A and B) showed that the risk of worsening cardiac structure (p<0.001) and function (p<0.001) was greatest in the severe subgroup, with hazard ratios (HRs) of 16.2 (95% CI, 6.0-44.1) and 9.0 (95% CI, 3.3-24.5), respectively. The risk of worsening cardiac structure (p=0.001) in the mild-moderate subgroup was higher than the normal subgroup, with hazard ratios (HRs) of 4.8 (95% CI, 1.9-12.5), and there was no significant difference in the follow-up of cardiac function between those two subgroups (p> 0.05). These findings suggest that the deterioration of cardiac structure and function was most apparent in the severe subgroup, despite treatments, and most cases experienced deterioration within the first 2 years from diagnosis.

Complications and death
TAK patients often presented complications such as infection (50%, 68/136), particularly respiratory infection (75%, 51/68), and 47.1% (24/51) of these patients used preventive anti-tuberculosis drugs during treatment. The dynamic changes in inflammation indicators were not associated with cardiac structure and function deterioration (p>0.05). During follow-up, 3 cases of death were recorded. One patient had severe aortic regurgitation and ventricular hypertrophy, and eventually died of coagulopathy, respiratory, and circulatory failure. The causes of death in the other 2 cases were aortic dissection and cerebrovascular accident.

Discussion
In this report, we provide a comprehensive description of cardiac structure and function changes in TAK patients in which we highlight the significance of early monitoring these changes in TAK patients with heart involvement and better management during treatment. The major advantage of our research is that it is based on a relatively large sample, providing a prospective study to investigate long-term cardiac outcomes and clarify risk factors for adverse cardiac structure and function in TAK population.
While previous study indicated that 8.6%-39.9% of TAK patients suffered from cardiac insufficiency [5,6], 60.1% of patients in our cohort presented cardiac structure lesions, from which it can be concluded that they are not rare for Chinese TAK patients. For patients with cardiac insufficiency, their echocardiography often manifests as valvular regurgitation and cardiac chamber enlargement, particularly on aortic valve and left ventricle. In our study, nearly half of TAK patients were found to have cardiac dysfunction, primarily aortic regurgitation and left ventricular enlargement. Cardiac structure abnormalities are caused by various factors. Inflammation in TAK often starts from the aortic root, which may explain the high proportion of aortic valve lesions. And aortic valve regurgitation causes left cardiac chamber enlargement or eventually heart failure [18]. In our study, the percentage of TAK patients exhibiting coronary involvement was approximately 6.1% compared with previous finding of 10-30% [19]. Coronary wall inflammation can lead to myocardial ischemia, which further accelerates early coronary atherosclerosis. Secondary myocardial ischemia caused by stenosis and occlusion of coronary arteries is also a significant cause for death of TAK [20].
It is very necessary to monitor cardiac structure and function changes in TAK patients with heart involvement as cardiac complications are known as major causes for low quality of life for these patients [12]. During our research, the patients were classified into 3 subgroups according to the severity of their cardiac structure and function impairment. A total of 136 patients (68.7%) were followed up, and 31.3% of patients review echocardiography again because of continued treatment at the local hospital. By investigating the long-term outcomes of those who underwent more than two echocardiographic examinations, we found that the deterioration of cardiac structure and function occurred in all subgroups, but the highest incidence was in the severe subgroup. Additionally, although some patients did not present any symptoms related to cardiovascular system, this does not mean that these symptoms will not appear in the future. There were 3 deaths in our study related to cardiac insufficiency, aortic dissection and cardio-cerebrovascular events, which indicated that cardiac involvements and progression were closely related to poor prognosis of TAK patients.
According to cox regression analysis, serum TNF-α level (≥ 8.1 pg/ml), aortic regurgitation and ventricular hypertrophy at baseline were the risk factors boosting cardiac structure deterioration.
TNF-α is an important pro-inflammatory factor and plays an important role in tissue fibrosis [21]. TNFα induces macrophages to produce IL-12, which can promote the differentiation of CD4 + T lymphocytes to Th1 cells and activate NK cells [22]. It has been confirmed by previous studies to be a key factor in the pathogenesis of TAK [23]. In this sense, elevated levels of TNF-α would suggest high levels of inflammation and subsequent fibrosis. Furthermore, several studies revealed that TNF-α directly promoted pressure overload-induced left ventricular hypertrophy and hypoxia-stimulated cardiac fibrosis, which were major reasons for the deterioration of cardiac structure in TAK [24]. Apart from this, if patients have aortic regurgitation and ventricular hypertrophy in the early stage, their cardiac structure would be more likely to worsen. Interestingly, the deterioration rate in mild aortic regurgitation group was significantly higher than the other two which already developed to the stage of moderate to severe regurgitation (P = 0.02). Finally, age and longer duration is also a critical risk factor for cardiovascular disease [25].
Although the prescription rates of glucocorticoids and immunosuppressive agents were quite close among the 3 subgroups during the period of induction and stabilization,the usage of drugs on improving cardiac function was higher for the severe subgroup, particularly digoxin and diuretic drugs, which was similar to the treatment during follow-up. When it comes to the TNF-α antibody, only 1 of 5 patients who have taken this drug suffered from cardiac structural deterioration. Since it is not widely used in clinical practice, no significant difference has been observed, so for further studies, more TNF-α antibody use for a larger sample size is needed. In the study, some severe patients exhibited improvements in disease through surgery. Analysis of the survival function curve indicates that it is quite essential to have active intervention during the first 2 years, a time when the cardiac function of patients is considerably impaired, so as to achieve better therapeutic effects and prolonged survival. Furthermore, the respiratory tract infection was quite frequent in the study, and TAK patients' immunity is rather low because of immunosuppressive agents, so it is indispensable to timely prevent or treat this kind of comorbidity.
There may be some limitations in this study. Fist, auxiliary inspection tests (electrocardiogram, cardiac magnetic resonance, etc.) would need to be performed in order to evaluate cardiac function more precisely in the future study. Second, missing values in the present study may affect the results to some extent. Further research is needed to confirm the conclusion.

Conclusion
Heart involvement with TAK patients in echocardiography mainly manifests as pathologically valvular and atrioventricular abnormalities. To find out if there is heart involvement and progress, it's indispensable to regularly monitor the serum TNF-α level and cardiac structure especially aortic regurgitation and ventricular hypertrophy in TAK patients during the long-term follow-up.  ESR, erythrocyte sedimentation rate; hsCRP, hypersensitive C-reactive protein; IL-6, interleukin-6; and TNF-α, tumor necrosis factor-α.      Based on serum TNF-α level (≥8.1pg/ml), aortic regurgitation and ventricular hypertrophy at baseline, patients could be separated into three risk groups: high risk (red, 40%), moderate (yellow, between 20 and 40%) and low risk (green, 20%).